What naïve listeners can teach us about sound quality

Most people perceive the sound quality of a playback system unconsciously and intuitively. If this sound perception of typical "naïve" listeners is to be determined instrumentally, the instrumental result must correlate well with this sound perception. Listening tests, in which the test persons evaluated a wide choice of music and many different audio devices based on several parameters, are the basis for instrumental procedures. We found that naïve, inexperienced listeners perceive sound quality primarily based on three quality dimensions. The first of these dimensions we call "timbre." By this, we mean how close to the original the listeners evaluate the signal's frequency response (tonality) and temporal precision. The dimension "immersiveness " describes how realistically the sound field "fills" the (virtual) space. The "distortion" dimension describes how "clean," i.e., unaffected by harmful influences of electronic and electroacoustic components, the signal sounds. The weighted ratings of these quality dimensions (MOST for timbre, MOSI for immersiveness, and MOSD for distortion), expressed as Mean Opinion Scores (MOS), together determine the perceived overall sound quality (MOSO) of a playback system.

We transferred the results of these listening tests into MDAQS and applied them. The instrumental evaluation method that is now possible with it eliminates the need for hundreds of human test subjects. The evaluation becomes faster, more reliable, and more efficient. The results in numerical values are easy to understand even for non-experts and, above all, always comparable.